In this products liability action, Procter & Gamble appeals from the district court's1 judgment enforcing a jury verdict and award against it for $300,000. The action arose following Patricia Kehm's death from toxic shock syndrome (TSS). Plaintiffs-appellees, Mrs. Kehm's surviving husband and children (Kehms), alleged that Mrs. Kehm's use of Rely tampons, a product designed and manufactured by Procter & Gamble, proximately caused her death, and that the tampons were defective and unreasonably dangerous. The jury returned a verdict in the plaintiffs' favor, awarding them $300,000 in compensatory damages, but rejecting the Kehms' demand for punitive damages.

On appeal, Procter & Gamble seeks a new trial, alleging five principal instances of prejudicial trial court error. Procter & Gamble contends the court erred in: (1) admitting reports prepared by the Center for Disease Control and various state health departments because the reports constituted inadmissible hearsay; (2) refusing to instruct the jury on Procter & Gamble's defense that Mrs. Kehm had a unique or "idiosyncratic" susceptibility to TSS; (3) refusing to grant a limiting instruction regarding Procter & Gamble's withdrawal of Rely tampons from the market; (4) admitting, and submitting to the jury, evidence bearing on punitive damages, including information on Procter & Gamble's financial resources; and (5) permitting an expert witness for the Kehms to perform an in-court demonstration. In addition, Procter & Gamble contests several other evidentiary rulings, which it claims resulted in unfair prejudice and improperly influenced the verdict. Finally, Procter & Gamble alleges misconduct by plaintiffs' counsel throughout the trial, which misconduct Procter & Gamble contends gave rise to an atmosphere of prejudice that precluded a fair trial. Procter & Gamble does not contest on appeal the sufficiency of the evidence to support the jury verdict on liability and causation.

Upon review of the entire record, we affirm the judgment of the district court.

I. Background.

Rely tampons consist of two absorbent materials contained in a polyester bag, which is, in turn, enclosed in a plastic inserter. The two materials, carboxymethylcellulose (CMC) and polyester, are not used in other tampons. Procter & Gamble first test-marketed Rely in 1974. By the end of 1979, the company was marketing the product nationally and had sold over half a billion Rely tampons.

Toxic shock syndrome was first identified and named in a November 1978 article by Dr. James K. Todd of the University of Colorado. Dr. Todd listed as symptoms of TSS fever, vomiting, diarrhea, low blood pressure, rash, and subsequent skin peeling. Dr. Todd hypothesized that Staphylococcus aureus (Staph A), a bacterium, caused the symptoms. Dr. Todd had observed the disease only in children, but by early 1980, several state health departments had reported TSS in adult women. In May 1980, the federal Center for Disease Control (CDC) published a summary of reported cases which indicated a strong correlation between the disease and menstruation. A June 27 CDC report drew the initial link between TSS and tampon use, but observed that the incidence of TSS did not appear to vary with the brand of tampon used. The CDC said the risk of TSS was too low to warrant a general recommendation against tampon use, but that women who were concerned about the disease could cut the risk by using tampons through only part of each menstrual period.

On August 8, 1980, Procter & Gamble received a responsible report that a Rely user had died of TSS. Procter & Gamble learned on August 21 of a Minnesota public health study showing twice as high a percentage of Rely users among TSS victims as among a control group of tampon-using women. On September 19, the CDC reported that its own study confirmed the Minnesota finding: Rely users were at a higher risk for TSS than users of other tampons. Procter & Gamble withdrew Rely from the market on September 22.

Patricia Kehm died on September 6, 1980, of what doctors later concluded was TSS. She had begun using Rely tampons on September 2, and had noticed the first symptoms of illness on September 3.

At trial, the Kehms advanced two theories of liability against Procter & Gamble: (1) Rely tampons were defective and unreasonably dangerous, because the CMC chips they contained fostered the growth of the bacterium that causes TSS; and (2) Procter & Gamble failed to warn Rely users of this defect when it knew or should have known of its existence. Procter & Gamble argued that it had no duty to warn because the tampons were not defective or unreasonably dangerous. Procter & Gamble also contended that Mrs. Kehm did not have TSS, and that even if she did, her unique susceptibility to the illness, rather than her use of Rely, was to blame for her death.

The jury awarded the Kehms $300,000 in compensatory damages, but no punitive damages. The district court denied Procter & Gamble's motions for a new trial and for judgment notwithstanding the verdict, and Procter & Gamble appealed.

II. Discussion.

A. Epidemiological Studies.

The district court admitted into evidence reports of epidemiological studies conducted by the CDC and various state health departments, and allowed the plaintiffs' expert, Dr. Bruce Dan, a former member of the CDC task force, to testify about these studies. Each of the studies analyzed the statistical relationship between tampon use and the incidence of TSS. The CDC's own study focused on cases of TSS which patients, doctors, and state health officials had reported to the CDC. The CDC administered a questionnaire both to a selected group of these TSS victims and to a control group. The questionnaire asked both groups about their exposure to and use of tampons, among other things. The CDC reported the results of its study, as well as the results obtained by the state agencies in their studies, in the CDC Morbidity and Mortality Weekly Report (MMWR). The state studies, including separate studies by health agencies of Utah, Wisconsin, and Minnesota, and a joint study by health authorities of Minnesota, Wisconsin and Iowa, were conducted in much the same manner as the CDC study.

Procter & Gamble contends that the MMWR reports of these studies, and much of Dr. Dan's testimony concerning them, are inadmissible as hearsay. Dr. Dan did not participate in the state studies, nor were either the interviewers who actually administered the CDC and state questionnaires to physicians, patients, and parents of patients, or the questionees themselves, available to testify.2

The district court admitted the MMWR reports under Fed.R.Evid. 803(8)(C), an exception to the hearsay rule which authorizes admission of investigative or "evaluative" public records and reports. Rule 803(8)(C) excepts from the hearsay rule, in civil cases, the following:

Records, reports, statements, or data compilations, in any form, of public offices or agencies, setting forth * * * factual findings resulting from an investigation made pursuant to authority granted by law, unless the sources of information or other circumstances indicate lack of trustworthiness.

Procter & Gamble argues that the MMWR reports are inadmissible because: (1) the statements they contain are not "factual findings" within the meaning of the rule; (2) the federal and state officials who prepared the studies did not have first-hand knowledge of the matters asserted; and (3) the statements are not trustworthy.

The public records and reports exception rests on "the assumption that a public official will perform his duty properly and the unlikelihood that he will remember details independently of the record." Fed.R.Evid. 803(8) advisory committee note, citing Wong Wing Foo v. McGrath, 196 F.2d 120 (9th Cir.1952). The rule "assumes admissibility in the first instance but with ample provision for escape if sufficient negative factors are present." Fed.R.Evid. 803(8) advisory committee note. The burden is on the party opposing admission to prove the report's untrustworthiness. See Baker v. Elcona Homes Corp., 588 F.2d 551, 558 (6th Cir.1978), cert. denied, 441 U.S. 933, 99 S. Ct. 2054, 60 L. Ed. 2d 661 (1979).

Procter & Gamble argues that the studies do not constitute "factual findings" under Rule 803(8)(C) because they rest on medical opinions and diagnoses, and because the government officials conducting the studies did not have first-hand knowledge of the data they collected. But as the district court observed, courts construing the term "factual findings" in Rule 803(8)(C) have given it broad scope. See, e.g., United States v. American Telephone & Telegraph Company, 498 F. Supp. 353, 360 (D.D.C.1980) (citing cases). They have often admitted government reports setting forth agency opinions and conclusions on the ground that such reports, because they are public records based on investigations conducted pursuant to lawful authority, are presumptively reliable. That is, "there is no reason not to admit the findings simply because they tend towards the conclusory rather than the factual end, unless, as Rule 803(8)(C) further provides, the 'sources of information or other circumstances indicate lack of trustworthiness.' " Id. See also Baker v. Elcona Homes Corp., supra, 588 F.2d at 558; United States v. School District of Ferndale, Michigan, 577 F.2d 1339, 1354 (6th Cir.1978). We agree with the district court that once a report is conclusively shown to represent findings of a public agency made pursuant to an investigation authorized by law, the central inquiry becomes whether the report is trustworthy.

Procter & Gamble argues that the studies are untrustworthy because they reflect numerous statistical biases and that, because it was unable to cross-examine the sources interviewed for the studies, it could not determine what effect the biases had on the validity of the data. We reject this argument, because it altogether ignores the rationale behind the hearsay exceptions--that a statement which carries other indicia of reliability is admissible for precisely that reason, despite the declarant's unavailability for cross-examination. That is, a party which would exclude evidence falling within the apparent scope of one of the hearsay exceptions must make an affirmative showing of untrustworthiness, beyond the obvious fact that the declarant is not in court to testify. Cf. Zenith Radio Corp. v. Matsushita Elec. Ind. Co., 505 F. Supp. 1125, 1148 (E.D.Pa.1980) (fact that findings are based in part on hearsay or on confidential sources not divulged to opposing party "does not ipso facto render the findings so untrustworthy as to be inadmissible"). Indeed, if this court were to accept Procter & Gamble's argument, it would have to hold public reports such as these inadmissible in every case where the opposing party had not had the chance to examine every person in the statistical pool the report covers.3 We refuse to give the 803(8)(C) exception so narrow a scope.

Moreover, though Procter & Gamble could not cross-examine individual patients and interviewers, it presented expert testimony of its own challenging the methodology of the government reports, and evidence rebutting the conclusions of those reports. It thus had ample opportunity to attack the probative value of the MMWR reports. The jury was therefore fully aware of the parties' conflicting assessments of the reports and, we believe, fully capable of evaluating the evidence on both sides. Cf. Migliorini v. United States, 521 F. Supp. 1210, 1218 (M.D.Fla.1981) (reports on CDC and Michigan studies, offered to prove length of period following swine flu vaccinations during which increased risk of contracting Guillain Barre Syndrome existed, admitted along with testimony criticizing the CDC study as utilizing too short a surveillance period and testimony that the CDC study was probably biased in favor of vaccinated population).

In Robbins v. Whelan, 653 F.2d 47 (1st Cir.), cert. denied, 454 U.S. 1123, 102 S. Ct. 972, 71 L. Ed. 2d 110 (1981), an automobile accident case, the First Circuit held it reversible error to exclude a report by the Department of Transportation's National Safety Bureau containing information on the maximum stopping distances for specific makes of automobiles. Id. at 49. The plaintiff had offered the report to support his argument that the defendant had been driving faster than the defendant claimed. The First Circuit determined that the report should have been admitted under Rule 803(8)(C) as a public agency "data compilation" setting forth the "purely factual" findings of "a detailed inquiry that the agency undertook." Id. at 50. The court presumed the report trustworthy because its purpose was to inform consumers, it did not relate to any litigation, and its motivation was unbiased. Id. at 50-51. The court rejected the argument that auto manufacturers who supplied some of the data might have overstated the performance of their products, as insufficient to overcome the burden of proving untrustworthiness. Id. at 51.

In the case at hand, the district court found that the CDC and state case studies employed procedures and methods widely accepted in the field of epidemiology. Procter & Gamble does not dispute this finding. Dr. Dan, as well as Procter & Gamble's own expert Dr. Feinstein, testified that epidemiologists regularly rely on studies of this kind. Furthermore, we note the timeliness of the investigations, the special skill of the agencies conducting them, and their lack of any motive for conducting the studies other than to inform the public fairly and adequately. See Rule 803(8)(C) advisory committee note (timeliness, investigative skill, and motive probative of trustworthiness or lack thereof); Baker v. Elcona Homes Corp., supra, 588 F.2d at 558 (police report trustworthy and admissible as public record where police officer, who possessed special skill in investigating auto accidents, began investigation minutes after accident, and there was no indication of an improper motive for the report); Reyes v. Wyeth Laboratories, 498 F.2d 1264, 1270 n. 1 (5th Cir.) (CDC studies admitted to show number of reported incidences of polio), cert. denied, 419 U.S. 1096, 95 S. Ct. 687, 42 L. Ed. 2d 688 (1974); cf. Givens v. Lederle, 556 F.2d 1341, 1346 (5th Cir.1977) (CDC studies admitted to show recognition by medical profession of occurrence of vaccineinduced polio where editor of report was present and subject to cross-examination).

Taking into account both the rationale for the public records hearsay exception and relevant precedent, we affirm the district court's admission of the MMWR reports of CDC and state agency epidemiological studies under Rule 803(8)(C). In light of the opportunity Procter & Gamble had to challenge the methodology and findings of the studies through the presentation of its own evidence, we also affirm the district court's decision not to exclude the reports as prejudicial under Fed.R.Evid. 403.

B. Defense of Unique Susceptibility.

Procter & Gamble argues that the trial court erred in refusing to instruct the jury to find for Procter & Gamble if the jury determined that Mrs. Kehm's death resulted from her unique susceptibility to TSS. Procter & Gamble argues that Iowa law, which applies in this diversity case, does not hold a manufacturer liable for injury resulting from the use of an ordinarily safe product where the injury occurs because of some unusual susceptibility of the user, not common to the general public, which the manufacturer could not reasonably have foreseen.

The Iowa case on which Procter & Gamble relies, Bonowski v. Revlon, Inc., 251 Iowa 141, 100 N.W.2d 5 (1959), is inapposite. There, the Iowa Supreme Court refused to hold that the manufacturer had been negligent in failing to foresee a user's idiosyncratic allergic reaction to a suntan lotion, or that it had breached any warranty, express or implied. But in 1970, eleven years after Bonowski, the Iowa court adopted the rule of strict products liability. See Hawkeye-Security Insurance Co. v. Ford Motor Co., 174 N.W.2d 672. Procter & Gamble does not dispute that that rule applies here. Under a strict liability regime, the manufacturer can be held liable even though it was not negligent and breached no warranty. We are therefore unable to see how Bonowski's holding that failure to foresee an idiosyncratic reaction does not constitute negligence or breach of warranty applies in this case, where neither negligence nor breach of warranty is at issue.

In this case, the focus is rather on whether Procter & Gamble had an obligation to warn users of Rely of the possibility of TSS (assuming, of course, that the product was not so dangerous that even due warning to users would not save the manufacturer from liability). In failure-to-warn cases, as in strict liability cases generally, liability does not turn on whether the risk of harm runs to a substantial number of persons. Rather, in determining whether a manufacturer has a duty to warn, courts inquire whether the manufacturer knew that there were even a relatively few persons who could not use its product without serious injury, and whether a proper warning would have helped prevent harm to them. See Reyes v. Wyeth Laboratories, supra, 498 F.2d at 1278; Basko v. Sterling Drug, Inc., 416 F.2d 417, 430 (2nd Cir.1969); Wright v. Carter Products, 244 F.2d 53, 58 (2nd Cir.1957). The plaintiffs introduced evidence to show that Procter & Gamble had reason to know of the TSS risk to some Rely users sometime before Mrs. Kehm used the product and died. Nor is this a case in which the plaintiff, unaware of her unusual susceptibility, suffered an adverse reaction so immediate and deadly that a proper warning could not have prevented harm. Mrs. Kehm used Rely tampons beginning four days before her death. Although she became increasingly ill during those four days, she continued to use Rely, unaware of its role in her illness. On these facts, the jury could reasonably find that, had a proper warning been given, Mrs. Kehm might have recognized her symptoms and discontinued using tampons.

Even if Mrs. Kehm was unusually susceptible, her susceptibility is not by itself necessarily enough to absolve Procter & Gamble of liability. The real issue is whether it breached its duty to warn. Thus the district court did not err in refusing to instruct the jury that if it found Mrs. Kehm unusually susceptible, it must for that reason alone find Procter & Gamble not liable.

C. Withdrawal of Rely from the Market.

Procter & Gamble contests the trial court's refusal to instruct the jury to consider the withdrawal of Rely from the market solely as background information. Procter & Gamble argues that the withdrawal of Rely from the market was a subsequent remedial measure which, under Rule 407 of the Federal Rules of Evidence, cannot be considered as evidence of negligence or culpable conduct.4 The trial court denied Procter & Gamble's request for the instruction, following Eighth Circuit decisions indicating that the rationale of Rule 407 does not apply in strict liability cases, where by definition negligence is not an issue. See Farner v. Paccar, Inc., 562 F.2d 518, 528 (8th Cir.1977); Robbins v. Farmers Union Grain Terminal Assoc., 552 F.2d 788, 792-93 (8th Cir.1977).

Since the trial court's decision in this case, this court has refined and qualified its position that negligence-like considerations do not come into play in products liability cases. In DeLuryea v. Winthrop Laboratories, 697 F.2d 222, 227-29 (8th Cir.1983), a case involving a drug manufacturer's duty to warn users of its products and the admissibility of a subsequent change in the warning statement the manufacturer packaged with the drug, we observed that some duty-to-warn cases raise issues of reasonableness and foreseeability closely akin to those present in negligence cases. Following the Fourth Circuit's analysis in Werner v. Upjohn Co., Inc., 628 F.2d 848, 858 (1980), cert. denied, 449 U.S. 1080, 101 S. Ct. 862, 66 L. Ed. 2d 804 (1981), we noted that whereas in negligence cases the inquiry focuses on the reasonableness of the defendant's conduct, in typical products liability cases the inquiry focuses not on the defendant's conduct but on the safety of the product. But where the product is inherently unavoidably unsafe, liability hinges on the adequacy of the warning to users, an issue which, like negligence in non-products cases, turns on the reasonableness of the defendant's responses to foreseeable dangers. DeLuryea, 697 F.2d at 228-29. Accordingly, we held that the rationale of Rule 407 applied with equal force under the circumstances of that case as in negligence cases, and therefore required the exclusion of evidence of subsequent changes in the drug warning leaflets. Id. at 229.5

Though we agree with and will in appropriate cases follow the DeLuryea reasoning, we do not believe it requires us to reverse the trial court in this case. For this case presents a slightly different issue than DeLuryea. Unlike the DeLuryea defendant, Procter & Gamble did not argue that the trial court should altogether exclude evidence of its withdrawal of Rely from the market; instead, it introduced the evidence on its own and requested that the jury be instructed to consider the evidence only for a specific limited purpose. Thus the issue before us is not whether the trial court erroneously admitted the evidence, but whether it abused its discretion in refusing to give a limiting instruction. We think it did not.

Initially, we note that even assuming Procter & Gamble was entitled to a limiting instruction, it was not entitled to the instruction it requested. Its instruction would have allowed the jury to consider evidence of the withdrawal only as "background" information. But Rule 407 allows the admission of evidence of subsequent remedial measures when offered for any purpose other than to show negligence or culpable conduct. The advisory committee notes to Rule 407 list, among the other purposes for which such evidence is admissible, admission to show the existence of a duty, if such existence is controverted.

Procter & Gamble introduced evidence of the Rely market withdrawal to show the company's good faith towards the consuming public. It attempted to portray the withdrawal as wholly voluntary on its part.6 The Kehms, on the other hand, attempted to show that Procter & Gamble withdrew Rely only after the Food and Drug Administration advised it that a soon-to-be-released CDC study showed a higher correlation between Rely and TSS than between other tampons and TSS, and after the FDA recommended that Procter & Gamble remove Rely from the market and notify users of the risks. The Kehms also introduced evidence to show that before it withdrew the product, Procter & Gamble requested that the CDC not mention Rely by name in the forthcoming study. Finally, the Kehms introduced evidence that Procter & Gamble had entered into a consent agreement with the FDA, in which it promised to withdraw Rely and inform consumers of the recent CDC study results.

Whether Procter & Gamble withdrew Rely voluntarily was hotly contested by the parties. At least four times during the trial, Procter & Gamble called attention to the voluntary nature of its withdrawal. Evidence relating to the withdrawal was thus admissible for purposes of impeachment, and perhaps also to show the existence or non-existence of a duty to withdraw the product. Clearly, then, Procter & Gamble was not entitled to an instruction that the jury consider the evidence for no purpose other than as an illustration of the case's "background."

Most important, given the substantial amounts of time and testimony devoted at trial to the circumstances surrounding Rely's market withdrawal, the likelihood that a limiting instruction would have lessened the possibility of prejudice, or that the absence of such an instruction would have generated prejudice, is very small. Indeed, a limiting instruction may well have served only to confuse the issues. The withdrawal of Rely from the market is inextricably bound up with Procter & Gamble's credibility and its portrait of itself as a corporation concerned for the public's welfare. It would only obscure the issues to instruct the jury to regard the fact of withdrawal as mere "background," when Procter & Gamble itself made the voluntariness of the withdrawal a part of its attempt to show that it acted responsibly during the TSS crisis.

Accordingly, we conclude that the district court did not abuse its discretion in refusing to give the jury instruction Procter & Gamble requested.7

D. Punitive Damages.

Prior to trial, Procter & Gamble filed a motion in limine to bar any reference by the plaintiffs to their claim for punitive damages, or to Procter & Gamble's financial resources, unless and until the court ruled that the Kehms had made out a prima facie case for punitive damages. The district court denied the motion, and on the fifth day of trial, ruled that the Kehms had made out a prima facie case for punitive damages. The trial court also denied Procter & Gamble's subsequent motions for a directed verdict on the punitive damages issue.

We hold on the record before us that the plaintiffs did not make out a case for punitive damages. Iowa permits an award of punitive damages only if the defendant's conduct was wanton or reckless, exhibiting complete disregard for the plaintiffs' rights. Feeney v. Scott County, 290 N.W.2d 885, 892 (Iowa 1980); McCarthy v. J.P. Cullen & Son Corp., 199 N.W.2d 362, 368-69 (Iowa 1972). Less than three months elapsed between the time that public health studies first suggested an association between TSS and tampon use and the time that Procter & Gamble removed Rely from the market. Procter & Gamble first learned of the forthcoming CDC study linking Rely to TSS in mid-September of 1980, after Mrs. Kehm's death on September 6. Given the voluminous evidence produced at trial regarding the speculation and uncertainty during that period as to the cause of TSS, Procter & Gamble's conduct cannot be labeled reckless.

Even if the Kehms produced enough evidence to make out a prima facie case,8 the record shows that by the end of the trial, the Kehms had failed to demonstrate Procter & Gamble's "complete disregard" for Mrs. Kehm's rights. In cases such as this, where there can be "reasonable disagreement over the relative danger and utility of an act," and where it cannot be said that the defendant acted with "subjective awareness of the consequences," the issue of punitive damages should not go to the jury. See Ellis, Fairness and Efficiency in the Law of Punitive Damages, 56 S.Cal.L.Rev. 1, 23 (1982). Accordingly, the trial court abused its discretion in allowing the Kehms' claim for punitive damages.

Notwithstanding our conclusion that the trial court erred in submitting the punitive damages question to the jury, we think no prejudice to Procter & Gamble resulted from the admission, in connection with the punitive damages issue, of evidence of its financial resources. Procter & Gamble's principal allegation of prejudice rests on the admission of its July 31, 1981 annual report, which contained a reference to a $75 million charge against earnings reflecting losses from the withdrawal of Rely from the market. Procter & Gamble argues that this entry unduly emphasized the disparity in wealth between the parties and created a "Robin Hood" state of mind in the jurors. But the record reveals that the trial court admitted nothing pertaining to the $75 million charge apart from the consolidated balance sheet itself. At Procter & Gamble's request, the trial judge excluded the portion of the annual report that specifically discussed the reasons for the $75 million charge, and directed the Kehms' counsel not to question the witnesses concerning that particular entry on the balance sheet.

Procter & Gamble, moreover, is a well-known national company. It is highly probable that the members of the jury were already aware of the disparities in wealth between the parties, and that they would have surmised that Procter & Gamble lost money by withdrawing Rely from the market. Under the circumstances, we believe the trial court's limited admission of the financial information adequately protected Procter & Gamble against whatever prejudicial effect the information could have had.

Finally, we note that the jury refused to award the Kehms punitive damages, and that the actual damages awarded fell within the zone of reasonableness. On the record before us, we cannot conclude that the trial court's submission of the punitive damages evidence to the jury requires reversal of the judgment.

E. In-Court Demonstration.

During the trial, the court permitted one of the Kehms' experts, Dr. Philip M. Tierno, to perform an in-court demonstration. Dr. Tierno testified that CMC, a component in Rely tampons, can interact with an enzyme found in the vagina, betaglucosidase, to create hospitable conditions for the growth of the Staph A bacteria that produce the TSS toxin. Dr. Tierno illustrated his point in the courtroom by combining, in distilled water, some betaglucosidase with the CMC chips from a Rely tampon.

Procter & Gamble contends that the trial court should not have permitted this demonstration, because, it says, Dr. Tierno did not combine the experimental substances in an environment that accurately replicates that of the human vagina. Procter & Gamble argues that even if the demonstration was relevant, it should not have been allowed because it confused and unduly prejudiced the jury.

Admissibility of demonstration evidence depends upon a foundational showing of substantial similarity between the tests conducted and what they purport to represent. Ramseyer v. General Motors Corp., 417 F.2d 859, 864 (8th Cir.1969). Procter & Gamble argues that Dr. Tierno's experiment did not meet this foundational requirement because the temperature in the courtroom was different from that in the vagina; he used betaglucosidase obtained from almond pits rather than from the vagina; he placed the enzyme in water rather than in menstrual fluid; and none of the other bacteria, microorganisms, or other substances normally present in the vagina were present in the experimental medium.

The trial court rejected Procter & Gamble's arguments on the ground that Dr. Tierno's testimony sufficiently explained the differences between the experimental environment and the actual vaginal environment to allow the jury to reach a reasoned conclusion as to their significance. The court also ruled that the probative value of the demonstration outweighed the danger of prejudice, in view of the ample opportunity Procter & Gamble had to cross-examine Dr. Tierno and rebut his testimony. See Fed.R.Evid. 403.

The decision to admit or exclude demonstration evidence rests largely in the discretion of the trial judge. A trial court should use caution in permitting such demonstrations, but the trial court's decision will not be overturned on appeal absent a clear showing of abuse of discretion. Ramseyer v. General Motors Corp., supra, 417 F.2d at 864. Perfect identity between experimental and actual conditions is neither attainable nor required. Id. Dissimilarities affect the weight of the evidence, not its admissibility. Id.

In the case at hand, Dr. Tierno testified that the lower courtroom temperature would serve merely to slow the speed of the reaction. He testified that the absence of various vaginal substances from the experimental medium would not affect the desired reaction, and that it was preferred practice in the scientific community to use almond-derived betaglucosidase whenever that enzyme was needed for experimental purposes. We think this testimony established the requisite foundational similarity between the experiment and the actual phenomena it purported to replicate. The trial court did not abuse its discretion in permitting the jury to consider the demonstration, for what it was worth, qualified as it was by both Procter & Gamble's extended cross-examination of Dr. Tierno and the testimony of Procter & Gamble's own expert witnesses. Under the circumstances, we cannot conclude that the demonstration resulted in unfair prejudice to Procter & Gamble. But cf. Randall v. Warnaco, Inc., Hirsch-Weis Div., 677 F.2d 1226, 1234 (8th Cir.1982) (warning of possibility of undue prejudice where jury saw filmed reenactment of tent fire in which plaintiff was injured).

F. Other Evidentiary Rulings.

Procter & Gamble also contests the following evidentiary rulings: the admission of two entries from Patricia Kehm's diary; the admission of evidence of other consumer complaints regarding Rely and the failure to give a limiting instruction thereon; and the exclusion of a letter from the CDC to a trade association, to which Procter & Gamble belonged, seeking funds for further study of TSS. We think none of these rulings amounted to reversible error.

1. Diary.

At trial, the court permitted Colleen Jones, Patricia Kehm's sister, to read two excerpts from Patricia's diary.9 The entries concerned normal domestic affairs and contained professions of Mrs. Kehm's love for her family. Procter & Gamble contends that the evidence should have been excluded, both as hearsay and as unduly prejudicial. The Kehms argue that the evidence was not hearsay and not cumulative. They say the diary entries were highly probative because they served to negate the inference that the trial testimony of family members, who expressed their admiration for Mrs. Kehm as a person and as a mother, portrayed Mrs. Kehm in "the best light possible." The trial judge admitted the evidence after examining the diaries outside the jury's presence and limiting the admission to the two entries. We do not think the trial court abused its discretion in so ruling.

2. Other Consumer Complaints.

Procter & Gamble objects to the trial court's admission of seven documents and the testimony of one witness, all relating to complaints Procter & Gamble received from consumers concerning Rely. Two of the documents were admitted only to show that Procter & Gamble had notice of the Rely-TSS link before Mrs. Kehm fell ill. The other five documents and the testimony of Susan Myers were admitted to show notice, causation, and Rely's "dangerousness."

As to the first two documents, one a letter from a consumer and the other a Procter & Gamble memorandum summarizing the complaints and inquiries it received concerning Rely in late August 1980, Procter & Gamble argues that neither document links Rely with any of the recognized symptoms of TSS. Consequently, Procter & Gamble contends, the documents were not admissible for even the limited purpose of showing notice. As to the other documents and the testimony of Susan Myers,10 Procter & Gamble contends that even if they were admissible to show notice, they were inadmissible to show causation in the Kehm case and therefore should have been the subject of a limiting instruction. We disagree.

Under Fed.R.Evid. 401, evidence of similar occurrences "might be relevant to the defendant's notice, magnitude of the danger involved, the defendant's ability to correct a known defect, the lack of safety for intended uses, * * * the standard of care, and causation." Ramos v. Liberty Mut. Ins. Co., 615 F.2d 334, 338-39 (5th Cir.1980), cert. denied, 449 U.S. 1112, 101 S. Ct. 921, 66 L. Ed. 2d 840 (1981). In this case, consumer complaints need not match the exact scientific description of TSS in order to show substantial similarity between other consumers' illnesses and Mrs. Kehm's illness. Procter & Gamble had ample opportunity, of which it availed itself, to rebut the force of the other complaints by pointing out dissimilarities between the complainers' symptoms and the symptoms of TSS. It was up to the jury to decide what weight to give the complaints from other consumers. See Henwood v. Chaney, 156 F.2d 392, 397 (8th Cir.) (weight to be given to evidence of prior accidents varies according to similarity of prior accidents to incident being tried), cert. denied, 329 U.S. 760, 67 S. Ct. 113, 91 L. Ed. 655 (1946). Accordingly, the trial court did not abuse its discretion in refusing to give the limiting instruction.

3. Defendant's Exhibit FP.

Procter & Gamble attempted at trial to introduce its exhibit FP, a letter from the CDC dated March 5, 1982, to the Health Industry Manufacturers Association, of which Procter & Gamble is a member, requesting funds from private industry for further TSS research. Procter & Gamble claims the letter was highly probative because it served to contradict the testimony of the plaintiffs' expert witness Dr. Dan. Procter & Gamble says that Dr. Dan testified that as of March 1982 the CDC already knew the cause of TSS and wished to conduct further research into the treatment and prevention of TSS, rather than into its cause. Procter & Gamble argues that the letter shows that even as of March 5, 1982, eighteen months after Mrs. Kehm died, the CDC had not concluded its research as to the cause of TSS, and that its solicitation of funds was for the purpose of continuing that research.

The district court excluded the letter on the grounds that it constituted hearsay; that Procter & Gamble had not shown a sufficient foundation to support the letter's admission under the business records exception to the hearsay rule (Fed.R.Evid. 803(6));11 that the letter did not serve to impeach Dr. Dan's testimony; and that the letter was irrelevant, cumulative, and unduly prejudicial.

The trial court has broad discretion in determining whether documents, otherwise admissible as business records, are sufficiently trustworthy. United States v. Page, 544 F.2d 982, 987 (8th Cir.1976). In order to lay a proper foundation for the admission of evidence under the business records exception to the hearsay rule, the proponent of the evidence must demonstrate that a document has been prepared and kept in the course of a regularly-conducted business activity. Here, the CDC letter was addressed not to Procter & Gamble but to a trade association to which Procter & Gamble belonged. Procter & Gamble offered to lay the foundation for the introduction of the document through the testimony of one of its own employees, who would testify, according to counsel, that Procter & Gamble "was informed [of the CDC's research plans] by the trade association, which forwarded a copy of the request for funds from the CDC" (emphasis added). Under these circumstances, it is questionable whether the letter satisfies the regularity condition that supplies the circumstantial indication of trustworthiness justifying excepting business records from the hearsay rule.12

The trial court also held the letter inadmissible as impeachment evidence. In the court's view, Dr. Dan had not testified that the CDC had by March 1982 resolved all questions concerning the cause of TSS. Therefore, the court reasoned, evidence purporting to impeach such testimony was not relevant. Having reviewed the record, we find nothing that contradicts the district court's assessment of Dr. Dan's testimony. Although Dr. Dan's statements regarding the cause of TSS support the inference that the CDC knew the cause by the time of trial, Dr. Dan did not testify that the CDC had discontinued its research into the disease's causes.

Finally, we observe that the trade association did not receive the letter or forward a copy to Procter & Gamble until some eighteen months after the commencement of the lawsuit. These circumstances increase the likelihood that the letter was untrustworthy. The decision whether to admit the letter was committed to the discretion of the district court, and on the record before us we cannot say that it abused that discretion.

G. Attorney Misconduct.

Procter & Gamble's final contention is that the conduct of the plaintiffs' attorney, Tom Riley, was so "inflammatory" that it incited passion and prejudice against Procter & Gamble, created a "circus" atmosphere, and deprived Procter & Gamble of a fair trial. Procter & Gamble points specifically to counsel's comments at trial regarding the company's financial resources, and to counsel's participation by telephone, while the trial was still in progress, in a nationally-broadcast television news program called "The Truth About Toxic Shock Syndrome."

In the context of our discussion of the punitive damages issue, see part IID supra, we addressed the possibility of prejudice arising from references to Procter & Gamble's financial resources. We concluded there that the likelihood of prejudice was slight, given the company's national prominence. As for counsel's participation in the television broadcast, the trial court's poll of the jury on the morning after the program appeared indicated that none of the members had seen the program or had heard of it. After trial, the court denied Procter & Gamble's motions for judgment notwithstanding the verdict and for a new trial, criticizing Mr. Riley's conduct but ruling that it did not prejudice Procter & Gamble's case.

After reviewing the record, and in light of all the circumstances of this case, we are convinced that Mr. Riley's conduct, though perhaps improper, did not incite jury passion and prejudice against Procter & Gamble. In particular, we note the jury's failure to award punitive damages, as well as the fact that it awarded $300,000 compensatory damages in face of a request by counsel for $5,000,000. Accordingly, we affirm the court's denial of a new trial.

III. Conclusion.

In conclusion, we note the difficult nature of this case. The suit required inquiry into highly technical subjects. TSS was a confusing disease that had previously been virtually unknown, and the case understandably evoked a high level of human emotion. The trial judge confronted a large number of difficult evidentiary questions. After reviewing the record as a whole, we are satisfied that both parties received a fair trial. Though we may not have made all the same rulings had we presided at trial, we cannot say that the trial judge abused his discretion or that any combination of his rulings unfairly prejudiced Procter & Gamble. Deference to the trial court's discretion is especially advisable where evidentiary rulings are concerned. It is, as well, the general policy of the Federal Rules to favor the admission of relevant evidence, absent an extremely strong showing that admission will lead to unfair prejudice. United States v. Dennis, 625 F.2d 782, 796-97 (8th Cir.1980); United States v. Day, 591 F.2d 861, 878 (D.C.Cir.1978).

Accordingly, we affirm the judgment of the district court.

LAY, Chief Judge, concurring.

I concur. However, I cannot agree with Judge Bright's analysis in Part II. C. concerning the admissibility of evidence of Procter & Gamble's withdrawal of Rely from the market as a post-remedial occurrence. This case was tried as a strict liability action; and consequently, Rule 407 of the Federal Rules of Evidence should have no applicability. Without reversing the trial court's admission of the evidence, the majority, however, approves a rationale previously adopted in DeLuryea v. Winthrop Laboratories, 697 F.2d 222 (8th Cir.1983) (Rule 407 can apply to a strict liability case). The question whether admission of Procter & Gamble's withdrawal of Rely from the market was proper, however, does not require us to approve the reasoning of DeLuryea.

In dicta, DeLuryea finds that in strict liability cases involving a drug manufacturer's duty-to-warn, Rule 407 applies and thus the admission of subsequent remedial conduct is limited. DeLuryea relies on the Fourth Circuit's decision in Werner v. Upjohn, Inc., 628 F.2d 848 (4th Cir.1980), cert. denied, 449 U.S. 1080, 101 S. Ct. 862, 66 L. Ed. 2d 804 (1981). With all due respect to our sister circuit, the Werner rationale, that "culpability" within Rule 407 was intended to apply to strict liability cases, is not historically or logically supported.13

In DeLuryea the panel distinguished strict liability cases in which the issue was the adequacy of the defendant drug manufacturer's warning from other strict liability actions. DeLuryea, 697 F.2d at 229 (citing Werner v. Upjohn, Inc., 628 F.2d 848 (4th Cir.1980), cert. denied, 449 U.S. 1080, 101 S. Ct. 862, 66 L. Ed. 2d 804 (1981)). In making the distinction, the DeLuryea panel attempted to distinguish Robbins v. Farmers Union Grain Terminal Assoc., 552 F.2d 788 (8th Cir.1977). In Robbins this court reasoned that Rule 407 is not applicable to strict liability cases, first, because the rule is addressed to issues of negligence or culpability, which by definition are irrelevant in a strict liability case; and, second, because the policy behind the exclusion is not persuasive in strict liability cases. Despite the unfortunate use of negligence terminology in jury instructions, there is a fundamental difference between strict liability actions and negligence actions. Robbins, 519 F.2d at 794-95.14 This distinction has been recognized in other Eighth Circuit decisions. See, e.g., Randall v. Warnaco, Inc., 677 F.2d 1226, 1232 (8th Cir.1982) (Bright, J.); Sterner v. U.S. Plywood-Champion Paper, Inc., 519 F.2d 1352, 1354 (8th Cir.1975).

The attempt in DeLuryea to distinguish Robbins is accomplished without a supporting rationale. I submit that the distinction, which creates an exception for duty to warn in drug cases, is without substance. It would be more appropriate to overrule Robbins than to generate variant applications of Rule 407 to issues that are identical. I believe DeLuryea should be limited to the facts of that case and its dicta should not become the controlling law within the circuit.

In concurring, I find the admission of evidence of Procter & Gamble's withdrawal of Rely from the market was proper. I do so for the reason that in a strict liability case involving a defective and unreasonably dangerous product, subsequent remedial change is relevant. The Seventh Circuit has made this clear:

[T]he plaintiff must establish that the "product in question has [not] lived up to the required standard of safety." [Citation omitted.] This, of course, requires proof that, inter alia: 1) the product as designed is incapable of preventing the injury complained of; 2) there existed an alternative design which would have prevented the injury; and 3) in terms of cost, practicality and technological possibility, the alternative design was feasible. Evidence of post-occurrence change which tended to satisfy plaintiff's burden on any of these issues would therefore, be relevant.

Lolie v. Ohio Brass Co., 502 F.2d 741, 744 (7th Cir.1974). A product can be made unreasonably dangerous when the manufacturer fails to provide safe warning to its users. Thus, evidence of Procter & Gamble's withdrawal was clearly admissible.